CN100334870C - Voice dialing system and realizing method - Google Patents
Voice dialing system and realizing method Download PDFInfo
- Publication number
- CN100334870C CN100334870C CNB2004100011178A CN200410001117A CN100334870C CN 100334870 C CN100334870 C CN 100334870C CN B2004100011178 A CNB2004100011178 A CN B2004100011178A CN 200410001117 A CN200410001117 A CN 200410001117A CN 100334870 C CN100334870 C CN 100334870C
- Authority
- CN
- China
- Prior art keywords
- user
- thresholding
- scene
- speech
- phonetic dialing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Abstract
The present invention provides a method and a system for realizing speech dialing. The method comprises the steps: receiving an incoming signal to trigger a speech prompt; answering by a user according to the speech prompt; recognizing the answer and outputting a corresponding confidence degree; confirming a processing scene according to the confidence degree to process the user answer and obtain a corresponding switched target terminal; switching the call to a target telephone according to the corresponding target terminal. The system comprises a storage device, at least one speech recognition server and a speech dialing control system, wherein the storage device is used for storing a speech library required to be identified and corresponding information. The speech recognition server is used for recognizing, according to the information of the storage device, speech input by the user according to the prompt. The speech dialing control system is used for controlling the running of the storage device and at least one speech recognition server which determines a switching target and a terminal according to the corresponding confidence degree of recognition results. The present invention has the advantages of distributed layout, low mistaken switching rate and easy maintenance.
Description
Technical field
The present invention relates to a voice dialing system, this system adopts distributed layout, and it is low that system misroutes rate, and is easy to safeguard.
Background technology
Along with the continuous maturation with the voice application technology be on the increase to various automations, intellectualizing system demand, variously finish the specific feature navigator interactive system of system based on voice suggestion guiding user and day by day increase, use and relate to phonetic dialing, mail, telephone number query, stock information, banking and other various information service fields.Therefore voice-based IAS becomes a very active field, and the phonetic dialing with extensive use demand then to be of this field use focus.
At present, the exchange of the most companies mode that adopts dial-in direct or receive calls by the manual operator.No matter which kind of mode all needs 20 seconds to 30 seconds time to search called subscriber's phone before each conversation at least, has hundreds of even thousands of staff's business for one, and the time that adds up is very surprising.And some adopt the telephone exchange of speech recognition technology, because the limitation of speech recognition technology and flow processing aspect etc. makes that correct switching rate, the hommization degree of system are not high.
In addition, can only transfer called subscriber's extension set of existing voice dialing system, when the called subscriber was not on the seat, existing telephone exchange system was just powerless.
In addition, existing voice dialing system generally can only simply be deployed on the machine, and the extendibility and the ease for maintenance of system are bad; The keeper holds and operates often more complicated, needs the special hilllock of special messenger.
Therefore, hope can have a kind of method and system, can have distributed layout, and it is low to misroute rate, and is easy to safeguard.
Summary of the invention
The objective of the invention is to overcome the above-mentioned shortcoming of prior art, for this reason, the invention provides a voice dialing system, what this system can distribute is deployed on many computers, also can be deployed on the computer simultaneously; The distinctive flow scheme design of this system can prevent wrong forward call to greatest extent; And the maintenance of information of this system is convenient and swift, does not need the special hilllock of special messenger.
In order to reach top purpose, technical scheme of the present invention is achieved in that
A kind of implementation method of phonetic dialing comprises step:
Receive incoming signal, to trigger the prompting of prompting user's voice;
The user replys with voice mode according to described voice suggestion;
Discern described replying, and export the confidence level of described recognition result correspondence;
Confirm to handle scene according to described confidence level,, obtain the terminal of the target correspondence of switching to handle described user answer;
Target phone is arrived in terminal relayed call according to described target correspondence; It is characterized in that: three kinds of confidence threshold value are set: first threshold value, second threshold value and the 3rd threshold value, their relation is: first threshold value＞second threshold value＞the 3rd threshold value;
The described step of handling scene according to described confidence level affirmation further comprises following content of operation: judge that whether described confidence level is greater than first thresholding, if greater than first thresholding, judge whether only to have a recognition result again, if only there is one, then change processing scene of the same name over to, confirm that the described switching target of identification is correct; If a plurality of recognition results are arranged, then change Nbest over to and handle scene, selecting one of them is described switching target; If less than first thresholding, then refuse to know judgement according to the 3rd thresholding, set up if refuse to know, require the user to re-enter voice; Be false if refuse to know, then enter and confirm to handle scene.
Described whether entering according to other threshold judgement confirms that the decision operation step of handling scene comprises: require user key-press to confirm recognition result, perhaps say name again, if the user confirms recognition result, then enter processing scene of the same name; If the user says name again, whether the confidence level of then judging new recognition result is greater than second thresholding, if greater than second thresholding, then judging whether only has a recognition result again, if only there is one, then change processing scene of the same name over to, confirm that the described switching target of identification is correct; If a plurality of recognition results are arranged, then change Nbest over to and handle scene, selecting one of them is described switching target; If less than second thresholding, require the user to re-enter speech.
The operation of described processing scene of the same name also comprises the steps: to judge whether the corresponding a plurality of terminals of the same name of described switching target, if a plurality of terminals of the same name are arranged, selecting one of them is the terminal of described correspondence; If only there is one, confirm as the terminal of described correspondence.
Described Nbest handles scene, and to select one of them be that the operating procedure of described switching target comprises: export a plurality of options, and the prompting user selects; The user selects according to described option.
When described option did not have right value, the user selected cancellation or phonetic entry again.
Also comprise the steps: if the terminal of described correspondence is not replied the mobile phone that forwarded user is predetermined.
Described first thresholding is that 70, the second thresholdings are 60.
The described operating procedure of knowing judgement according to refusing in other threshold judgement comprises: whether judge described confidence level less than the 3rd thresholding, if less than the 3rd thresholding, refuse to know and set up; If greater than the 3rd thresholding, confirm described switching target by the user; Wherein, described the 3rd thresholding is 45.
In order to reach top purpose, the present invention also provides a kind of system of phonetic dialing, is used to realize said method, it is characterized in that, described system comprises: storage device is used to store the sound bank and the corresponding informance of needs identification; At least one speech recognition server is used for the voice according to the prompting input according to the information Recognition user of described storage device; The phonetic dialing control system is used to control the operation of described storage device and at least one speech recognition server; Wherein, be provided with the process control module that to determine the flow process redirect between switching target and terminal and the associated scenario according to the confidence level of recognition result correspondence in the described speech recognition server.
This system also comprises:
Keeper's end is used between keeper and described phonetic dialing control system mutual;
Load balance system, be used for according to the load of described at least one speech recognition server of Instruction Selection of described phonetic dialing control system lighter one of carry out speech recognition.
Voice dialing system disclosed by the invention can distributedly be deployed on many machines; Method of the present invention can reduce the rate of misrouting to greatest extent; The interface of same time keeper end makes that whole system is easy to safeguard.
Description of drawings
Fig. 1 is the principle deployment diagram of system of the present invention;
Fig. 2 is the flow chart of the process module of the embodiment 1 of system of the present invention;
Fig. 3 is the flow chart of the process module of the embodiment 2 of system of the present invention;
Fig. 4 is the interface of system manager's end of the present invention to system monitoring;
Fig. 5 is the interface of system manager's end of the present invention to maintenance of information;
Fig. 6 plays the flow chart of steps of phonetic synthesis for the present invention.
Embodiment
In order to make those skilled in the art person understand the present invention better, the present invention is described in further detail below in conjunction with drawings and embodiments.
Fig. 1 is the deployment diagram of system of the present invention; Speech recognition server wherein, voice synthesizing server, load balance system, database, keeper's end, phonetic dialing control system etc. do not need to be deployed on the independent machine, can one or several system deployment on same machine, even whole voice dialing system is deployed on same the machine.But on same the machine a plurality of voice synthesizing servers can not be arranged, a plurality of speech recognition servers can not be arranged; When whole system has only a speech recognition server and a voice synthesizing server, then do not need load balance system, each phonetic dialing control system need be discerned or synthetic the time, directly link to each other with identification or synthesis server.
Fig. 2 is the flow chart of process control module of the embodiment 1 of native system.Confidence level wherein is the be sure of degree of speech recognition server to this recognition result, and in the time of confidence level＞70, general this recognition result of identified server is all very accurate; Recognition result 1, recognition result 2 etc. are that identified server sorts this recognition result according to confidence level, and coming top recognition result 1 is that confidence level is the highest, by that analogy.
So this flow process is at first carried out redirect according to the recognition result of home court scape,, at this time can directly forward processing scene of the same name to if the confidence level of recognition result 1＞70 in the time of and the confidence level of recognition result 2＜65, just show that recognition result 1 is very credible; Otherwise, change affirmation scene and telephone subscriber over to and check whether recognition result 1 is the people that the telephone subscriber will look for.
Equally, also be to carry out redirect when the telephone subscriber says name for the second time according to recognition result, if the confidence level of recognition result 1＞70, in the time of and the confidence level of recognition result 2＜65, just show that recognition result 1 is very credible, at this time can jump to processing scene of the same name; But then change Nbest (N best) scene when not satisfying this condition this time, that is: several candidate's recognition results are all listed, and power supply words user selects.
Distinctive processing capacity of the same name, when appearance was of the same name, the prompting telephone subscriber had a plurality of employees that are this name, and provides each employee's of the same name supplementary features, for example: department, sex etc.
When the employee not on the seat, and opened switching cell-phone function the time, system also can be transferred on this employee's the mobile phone for the telephone subscriber.
No matter be according to confidence level carry out scene hop, repeatedly confirm, many recognition results are handled, or processing of the same name, switching mobile phone, all be in order to reduce the rate of misrouting to greatest extent, reduce system interaction time and interaction times simultaneously, and can find called employee for the telephone subscriber as far as possible.
Fig. 3 is the flow chart of process control module of the embodiment 1 of native system.Wherein, fractional value 70 among the figure (first thresholding), 60 (second thresholdings) are just for an example, in addition also have one 45 (the 3rd thresholding--refuse field of awareness value, that is: be lower than this mark, the identification engine will no longer return mark, system does to refuse to know and handles, and no longer walks according to flow process), can adjust different identification engines, different application;
In this flow chart, do not list the situation of refusing to know (refusal identification), refuse the processing known such as following false code usually:
If (refuse continuously know number of times＜3)
Reenter current scene; Again play the signal language of current scene;
else
System changes the manual operator;
Owing to do not list the situation of no recognition result in this flow chart, when the user did not have phonetic entry or sound too small, system was considered as not having recognition result.Generally in the time of no recognition result, system can do does not have the input processing, and it is handled as following false code:
If (number of times＜3 of not having input continuously)
Reenter current scene; Again play the signal language of current scene;
else
System thinks that the user leaves, initiatively on-hook;
Red frame is respectively two algorithms among the figure, and in the use of home court scape is algorithm 1; In inferior scene use is algorithm 2:
1) algorithm 1:
If (all recognition results all＜45)
Handle as refusing to know
Else (more than identification mark＞70 are arranged)
Forward Nbest to and handle scene
Else (having only one＞70)
Forward processing scene of the same name to
Else//promptly: do not have mark) greater than 70
Enter the affirmation scene;
2) algorithm 2:
If (all recognition results all＜45)
As refusing to know processing (referring to 2)
Else (having only identification mark＞60)
Forward processing scene of the same name to;
Else//promptly: a plurality of identification marks are arranged greater than 60, perhaps do not have recognition result greater than 60)
Forward Nbest to and handle scene;
Wherein, confirm that scene is work like this:
System: XX, affirmation is please by " # " number key, otherwise says please whom you look for again
User: by " # " number key
Because the user is that button is represented to determine, system can identification error, so the score value of " determining " this recognition result is not processed; On the contrary, if user besides go out certain name, then system will discern mark to it and processes, and promptly algorithm 2;
It is that the highest several recognition results of system identification mark are processed that Nbest (n best) handles scene, and so-called Nbest is meant the setting of identification engine according to system, and n preferable recognition result returned in active user's input; Usually, this n preferable recognition result just is not higher than 45 (the 3rd thresholding--refuse field of awareness value), but is higher than 60 (second thresholdings); From flow chart, come as can be seen, when having only an identification mark＞60, just can skip this scene, otherwise (a plurality of identification marks＞60 are just arranged, and perhaps all identification marks are between 45 and 60) all can enter this scene.
This scene is work like this:
User: Zhang San
System: Zhang San is please by 1 key, opens the mountain please by 2 keys, cancellation please by "
*" number key,
User: by " 1 " number key
Because this scene requires the user key-press input, system can identification error, so this scene Recognition result's score value is not processed;
And processing scene of the same name is work like this:
User: Zhang San
System: have two to be Zhang San, the Zhang San of Beijing factory is please by 1 key, and research institute Zhang San is please by 2 keys, cancellation
Please by "
*" number key,
User: by " 1 " number key
Because this scene requires the user key-press input, system can identification error, so this scene Recognition result's score value is not processed;
Part after the forward call scene is optional part, and can transfer as required mobile phone or home phone number etc. can not have this part content yet.
In addition, this method can be applied to different speech recognition engines, for example: Nuance, ViaVoice etc.
Following table is to the statistics of the transit case of different thresholdings
| Staff number | First thresholding | Misroute rate | Need enter the percentage of confirming scene |
| 3000 | 60 | ＞10％ | 30％ |
| 65 | 5％ | 40％ | |
| 70 | 0.8％ | 50％ | |
| 75 | 0.1％ | 70％ | |
| 80 | Almost do not have | More than 90% |
From flow chart, come as can be seen, when having only an identification mark, just can skip this scene greater than the 2nd thresholding, otherwise (a plurality of identification marks are just arranged greater than the 2nd thresholding, perhaps all identification marks are between the 3rd thresholding and the 2nd thresholding), thus be not difficult to draw such conclusion: and the too high or low excessively percentage that all can cause entering the Nbest scene in inferior scene of the 2nd thresholding increases.
The too high meeting of the 3rd thresholding causes reject rate to increase; Cross and low can cause system to return a lot of incoherent recognition results, make availability reduce.This table is the statistics that headcount (that is: speech recognition vocabulary number) is done when being 3000.
Fig. 4 is the system monitoring interface of native system keeper end, by this interface, the keeper not only can watch system state information (port free time whether, system whether normal etc.), log information constantly (for example: internal memory operating position etc.), can also the On/Off system, system parameters, the conversation quantity of inquiry a period of time are set.
Fig. 5 is the system information maintenance interface of keeper's end of native system.By this interface, keeper's initialization phone directory and some database synchronization, increase/modification/certain employee's of deletion information, the information of increase/modification/certain department of deletion, compiling grammar file, modification signal language information etc.Wherein the result of the operations such as information of initialization phone directory and some database synchronization, increase/modification/certain employee's of deletion information, increase/modification/certain department of deletion is saved in the database by the phonetic dialing control system; Compiling grammar file wherein operates in the phonetic dialing control system and carries out, and the grammer bag after the compiling is updated in each speech recognition server; Operation such as modification signal language information wherein is by in the phonetic dialing control system database.
In general, system of the present invention comprises: load balance system, it is after the order of receiving voice dialing system identification or synthesizing, watch the load condition of current each identified server or synthesis server, find the lightest station server of present load, machine name is told to the phonetic dialing control system, make phonetic dialing control system and this identification or synthesis server connect.
At least one speech recognition server, this server will provide current state when load balance system is inquired about this server state: busy, not busy; If busy, also to provide and how long hurry.After this server and voice dialing system connected, its was responsible for concrete identifying operation, just disconnected then.The keeper can safeguard the information such as grammar file of speech recognition server by keeper's end system.
At least one voice synthesizing server, this server will provide current state when load balance system is inquired about this server state: busy, not busy; If busy, could be idle after how long also will providing.After this server and voice dialing system connected, its was responsible for concrete phonetic synthesis operation, just disconnects then.The keeper can safeguard the information such as recording file, sound storehouse of voice synthesizing server by keeper's end system.
At least one keeper's end system.Keeper's end system is safeguarded the grammar file of speech recognition system, the signal language of speech synthesis system, the information such as employee department in the database by voice dialing system.
A phonetic dialing control system, this system are responsible for voice plate card control, handle each recognition result, each synthetic content are set; This system is responsible for holding the keeper information of maintenance to be updated to respectively in phonetic dialing control system, each speech recognition server and the database by the keeper in addition.
A database is used to store employee's information such as telephone number, name, name phonetic, department; For the employee that duplication of name is arranged, also to store additional identification information and the phonetic thereof that to distinguish this employee.
Wherein, phonetic dialing control system links with keeper's end, database, load balance system respectively; Load balance system also links to each other with each voice synthesizing server, speech recognition server except linking with the phonetic dialing control system.
In the said system, the phonetic dialing control system also comprises flow process control unit, voice plate card and control unit thereof, grammer bag compiler etc.
Keeper's end also comprises voice dialing system state detection module and voice dialing system maintenance of information module.
A kind of method of distributed sound dialing, its job step is as follows:
A. the keeper monitors voice dialing system by keeper's end and safeguards;
B. when the telephone subscriber made a phone call to voice dialing system, the phonetic dialing control system was replied by voice plate card, and flow process position initial condition is set.
C. the phonetic dialing control system is provided with the content of signal language and sends to load balance system and will carry out the request of phonetic synthesis;
D. load balance system is inquired about the state of each voice synthesizing server, seeks the lightest one of load, tells the phonetic dialing control system its machine name;
E. the phonetic dialing control system is connected with this voice synthesizing server, the phonetic dialing control system connects to the delivery of content that will synthesize this voice synthesizing server, this voice synthesizing server connects a synthetic result and returns to the phonetic dialing control system, disconnect connection then, the state of oneself is put the spare time;
F. phonetic dialing control system play cuing language is provided with grammar file simultaneously and sends the request that will carry out speech recognition to load balance system;
G. load balance system is inquired about the state of each speech recognition server, seeks the lightest one of load, tells the phonetic dialing control system its machine name;
H. the phonetic dialing control system is connected with this speech recognition server, the phonetic dialing control system is wanting content identified to pass to this speech recognition server, this speech recognition server returns to the phonetic dialing control system to the result of identification, disconnect connection then, the state of oneself is put the spare time;
I. the phonetic dialing control system is carried out scene hop according to the recognition result that returns, and then jumps to c if not done state
A process control module is arranged in the phonetic dialing control system, and it decides the redirect of flow process according to the confidence level of recognition result and recognition result, can farthest reduce the rate of misrouting like this.
When certain employee is not on the seat, this employee's that can transfer mobile phone.
When whole system has only a speech recognition server and a voice synthesizing server, then do not need load balance system, each phonetic dialing control system need be discerned or synthetic the time, directly link to each other with identification or synthesis server.
Can several system deployment on same machine, even whole voice dialing system is deployed on same the machine.But on same the machine a plurality of voice synthesizing servers can not be arranged, a plurality of speech recognition servers can not be arranged.
Keeper end can safeguard the running status of voice dialing system, for example: whether occur unusual, have what user is online, nearest message accounting etc. (referring to accompanying drawing 3); Can also some parameters of system be provided with; Can also safeguard the employee information in the system, department information etc. simultaneously.
Though described the present invention by embodiment, those of ordinary skills know, the present invention has many distortion and variation and do not break away from spirit of the present invention, wish that appended claim comprises these distortion and variation and do not break away from spirit of the present invention.
Claims (10)
1, a kind of implementation method of phonetic dialing comprises step:
Receive incoming signal, to trigger the prompting of prompting user's voice;
The user replys with voice mode according to described voice suggestion;
Discern described replying, and export the confidence level of described recognition result correspondence;
Confirm to handle scene according to described confidence level,, obtain the terminal of the target correspondence of switching to handle described user answer;
Target phone is arrived in terminal relayed call according to described target correspondence; It is characterized in that: three kinds of confidence threshold value are set: first threshold value, second threshold value and the 3rd threshold value, their relation is: first threshold value＞second threshold value＞the 3rd threshold value;
The described step of handling scene according to described confidence level affirmation further comprises following content of operation: judge that whether described confidence level is greater than first thresholding, if greater than first thresholding, judge whether only to have a recognition result again, if only there is one, then change processing scene of the same name over to, confirm that the described switching target of identification is correct; If a plurality of recognition results are arranged, then change Nbest over to and handle scene, selecting one of them is described switching target; If less than first thresholding, then refuse to know judgement according to the 3rd thresholding, set up if refuse to know, require the user to re-enter speech; Be false if refuse to know, then enter and confirm to handle scene.
2, the implementation method of phonetic dialing as claimed in claim 1, it is characterized in that: described whether entering according to other threshold judgement confirms that the decision operation step of handling scene comprises: require user key-press to confirm recognition result, perhaps say name again, if the user confirms recognition result, then enter processing scene of the same name; If the user says name again, whether the confidence level of then judging new recognition result is greater than second thresholding, if greater than second thresholding, then judging whether only has a recognition result again, if only there is one, then change processing scene of the same name over to, confirm that the described switching target of identification is correct; If a plurality of recognition results are arranged, then change Nbest over to and handle scene, selecting one of them is described switching target; If less than second thresholding, require the user to re-enter speech.
3, the implementation method of phonetic dialing as claimed in claim 1 or 2, it is characterized in that: the operation of described processing scene of the same name also comprises the steps: to judge whether the corresponding a plurality of terminals of the same name of described switching target, if a plurality of terminals of the same name are arranged, selecting one of them is the terminal of described correspondence; If only there is one, confirm as the terminal of described correspondence.
4, the implementation method of phonetic dialing as claimed in claim 1 or 2 is characterized in that: described Nbest handles scene, and to select one of them be that the operating procedure of described switching target comprises: export a plurality of options, and the prompting user selects; The user selects according to described option.
5, the implementation method of phonetic dialing as claimed in claim 4 is characterized in that: when described option did not have right value, the user selected cancellation or phonetic entry again.
6, the implementation method of phonetic dialing as claimed in claim 3 is characterized in that: also comprise the steps: if the terminal of described correspondence is not replied the mobile phone that forwarded user is predetermined.
7, the implementation method of phonetic dialing as claimed in claim 1 or 2 is characterized in that: described first thresholding is that 70, the second thresholdings are 60.
8, the implementation method of phonetic dialing as claimed in claim 1 or 2, it is characterized in that: the described operating procedure of knowing judgement according to refusing in other threshold judgement comprises: judge that whether described confidence level is less than the 3rd thresholding, if less than the 3rd thresholding, refuse to know and set up; If greater than the 3rd thresholding, confirm described switching target by the user; Wherein, described the 3rd thresholding is 45.
9, a kind of system of phonetic dialing is used to realize the method as claim 1-8, it is characterized in that described system comprises: storage device is used to store the sound bank and the corresponding informance of needs identification; At least one speech recognition server is used for the voice according to the prompting input according to the information Recognition user of described storage device; The phonetic dialing control system is used to control the operation of described storage device and at least one speech recognition server; Wherein, be provided with the process control module that to determine the flow process redirect between switching target and terminal and the associated scenario according to the confidence level of recognition result correspondence in the described speech recognition server.
10, the system of phonetic dialing as claimed in claim 9 is characterized in that, this system also comprises:
Keeper's end is used between keeper and described phonetic dialing control system mutual;
Load balance system, be used for according to the load of described at least one speech recognition server of Instruction Selection of described phonetic dialing control system lighter one of carry out speech recognition.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2004100011178A CN100334870C (en) | 2004-01-20 | 2004-01-20 | Voice dialing system and realizing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2004100011178A CN100334870C (en) | 2004-01-20 | 2004-01-20 | Voice dialing system and realizing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1617554A CN1617554A (en) | 2005-05-18 |
| CN100334870C true CN100334870C (en) | 2007-08-29 |
Family
ID=34763036
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2004100011178A Expired - Fee Related CN100334870C (en) | 2004-01-20 | 2004-01-20 | Voice dialing system and realizing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100334870C (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101136962B (en) * | 2006-08-30 | 2011-06-08 | 许军 | Telephone switching system and switching method thereof |
| US8055502B2 (en) * | 2006-11-28 | 2011-11-08 | General Motors Llc | Voice dialing using a rejection reference |
| CN101426047B (en) * | 2007-10-31 | 2013-04-24 | 陈修志 | Intelligent voice control telephone |
| CN106875939A (en) * | 2017-01-13 | 2017-06-20 | 佛山市父母通智能机器人有限公司 | To the Chinese dialects voice recognition processing method and intelligent robot of wide fluctuations |
| JP6521397B2 (en) * | 2017-07-05 | 2019-05-29 | Ｎｅｃプラットフォームズ株式会社 | Telephone system, telephone, and telephone transfer method |
| CN111050002B (en) * | 2019-12-17 | 2021-07-30 | 北京鸿博信通科技有限公司 | Intelligent telephone exchange and working method and system thereof |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030016807A1 (en) * | 2001-07-23 | 2003-01-23 | Nec Corporation | Voice CODEC designating system |
-
2004
- 2004-01-20 CN CNB2004100011178A patent/CN100334870C/en not_active Expired - Fee Related
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030016807A1 (en) * | 2001-07-23 | 2003-01-23 | Nec Corporation | Voice CODEC designating system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1617554A (en) | 2005-05-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US6404876B1 (en) | System and method for voice activated dialing and routing under open access network control | |
| CN100407291C (en) | Dynamic and adaptive selection of vocabulary and acoustic models based on a call context for speech recognition | |
| US6487277B2 (en) | Apparatus and method for improving the user interface of integrated voice response systems | |
| US7184540B2 (en) | Personality based matching of callers to agents in a communication system | |
| CN101573750A (en) | Method for selecting interactive voice response modes using human voice detection analysis | |
| KR20060012601A (en) | System and method for automated customer feedback | |
| US20010012336A1 (en) | Automated message system for a calling party | |
| US20030152199A1 (en) | Dialogue device for call screening and Classification | |
| RU2006103855A (en) | PROVISION OF SERVICES AND TRANSACTIONS IN THE PHONE NETWORK | |
| WO2021135317A1 (en) | Method and system for making call on the basis of order | |
| CN107302639A (en) | A kind of method for initiating queuing customer service self-assisted voice interaction | |
| CN100334870C (en) | Voice dialing system and realizing method | |
| US20080008308A1 (en) | System and method for routing calls | |
| US7412386B2 (en) | Directory dialer name recognition | |
| US6418418B1 (en) | Transaction information processing system | |
| JP2010258556A (en) | Prank call management system | |
| EP1381210B1 (en) | Caller identification based on voice recognition | |
| CN101202795A (en) | Method and system for audio frequency content user recording | |
| US7149295B2 (en) | Technique for effectively controlling communication links to an information assistance service | |
| US20020154757A1 (en) | Method for call distribution in a call center and associated call distribution apparatus | |
| US20090248418A1 (en) | Speech Recognition and Statistics-Based Call Route Determination | |
| CN109862195A (en) | Intelligent sound calling system and method based on information automatic error-correcting | |
| US6678365B1 (en) | Telephone information service system | |
| EP1424844A1 (en) | Human fallback method and system for interactive voice response systems | |
| CN113890949A (en) | Intelligent voice outbound method and device for intention recognition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20070829 Termination date: 20210120 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |